It is frequently necessary to determine the behavior of a test item under motion and/or loading to see whether it is suitable for some purpose. As an example, it may be desirable to move a sample or device (e.g., a replacement part or implant for a human skeletal structure) in some particular pattern of translation and/or rotation to determine its behavior in conditions simulating actual use. In other cases, it may be desirable to expose a sample of some substance to loading to determine its strength. In yet other situations, it may be desirable to subject a test sample to both motion and loading to determine its behavior.
Testing is often done in uniaxial testing apparata, which generate motion and loads along a single axis of the test item. As an example, a uniaxial testing apparatus may fix portions of a test item to opposing mounts and then move the mounts apart or together to determine the test item's behavior under tension or compression. Alternatively, the uniaxial testing apparatus may rotate the mounts with respect to each other to determine the test item's behavior under torsion.
Uniaxial testing apparata are currently the most popular types of testing apparata in the marketplace. However, they have limited usefulness because under real-world conditions, almost every structure experiences multiaxial motion and loading rather than uniaxial motion and loading. Therefore, uniaxial testing apparata generally do not accurately simulate real-world conditions. It would therefore be desirable to have testing apparata available which allow motion and/or loading of a test item in more than one degree of freedom so that real-world conditions are reproduced. The above-named inventors do not know of any commercially available multiaxial testing apparata, but an example of a possible such system is illustrated in FIG. 1, which illustrates an articulated serial linkage robotic arm which is anchored to the testing environment at one end and which terminates in pincers at its other end. A test item is then anchored to a mount on the testing environment so that it may be grasped by the pincers, and actuation of the robotic arm along and about various axes can allow the test item to experience displacements and rotations about these axes so that its behavior may be observed. Additionally, if stress/strain measurements in the test item are desired, sensors on the robotic arm, the mount, and/or on the test item can allow measurements of the forces experienced by the test item.
Since serial-linkage robotic arms are widely sold for automation of industrial and research processes, such a multiaxial testing apparata would be relatively easy to construct. However, because serial linkages are not very stiff, this testing apparatus would usually not allow for great accuracy in pose control (i.e., control of its displacement and rotation about xyz axes) and force measurements: any play in the joints between the links, and any bending and/or strain in the links, could result in measurement and control errors which are difficult to compensate. Furthermore, accurate simulation of real-world motions often calls for the use of cyclical motions, shock (unit impulse) motions, and unit step motions, and most robotic arms are incapable of accommodating such dynamic behavior. As an example, most robotic arms cannot accommodate complex repetitive motions of any significant amplitude at frequencies around and below 1 Hz.
The aforementioned testing problems were noted by the above-named inventors in the context of testing sections of mammalian spinal sections under real-world motion and loading characteristics. Mammalian spines are formed of a number of vertebrae having different complex shapes, some of which are fused together and some of which are separated by flexible cartilage pads. In effect, the spine is itself a serial linkage. While spinal health is of great importance--most people experience back problems and pain during some period of their lifetimes--the behavior of the spine during various motions and loadings is not well understood. It was therefore desired to investigate the behavior of the spine (and sections thereof), as well as of proposed surgical corrective implants. However, it was quickly learned that uniaxial testing apparata would be of little use in developing an accurate model of spinal behavior under real-world three-dimensional motion and loading conditions. Further, no multiaxial testing apparata could be found which would allow for sufficiently accurate pose (motion) control (with errors of no more than 0.001 inches of displacement being desired), or which had sufficiently fast response times that it could, for example, accommodate cyclical motions on the order of 1 Hz (for example, to simulate spinal motion while running). The inventors therefore sought to develop a testing apparatus which would alleviate the aforementioned deficiencies.